风暴分类识别技术在人工防雹中的应用

利用新一代多普勒天气雷达资料,在风暴跟踪识别算法的基础上,发展了风暴分类技术,以提高人工防雹作业指挥的效率。首先以SCIT算法为基础,结合风暴的结构特征,综合利用雷达、探空资料,自动提取风暴结构特征指数;其次采用基于决策树模型的风暴自动分类技术,将风暴按强度分为雷雨云、单体风暴、多单体风暴和强风暴;最后根据风暴强度、高度和位置等属性,对有可能产生冰雹的单体,结合GIS,自动对下游方向或附近作业点进行预警或输出作业参数。通过对2006—2014年期间重庆、辽宁大连和河南三门峡三地发生的较为典型的31次冰雹天气过程、182站次冰雹样本的检验来看:该方法通过对风暴按强度、垂直结构等综合属性进行分类,能有效提高冰雹识别的命中率、降低空报率,其中强风暴的命中率能达到100%,空报率仅为11.4%。能有效提高人工防雹作业的自动化程度,对防雹作业的科学决策有着重要参考作用。

Application of Storm Auto Classification Technology in Artificial Hail Prevention

To increase the efficiency in artificial hail suppression operation, a storm auto classification technology is developed based on the storm tracking and recognition algorithm using the new generation Doppler weather radar data in this study. The characteristic indices of storm structure are firstly automatically extracted by using the radar and sounding data and the SCIT algorithm. And then according to the intensities, the storms are classified into weak thunderstorm, single cell storm, multi cell storm and severe storm by adopting the automatic classification technology of decision tree model. Finally, the early warnings on the downstream direction of the storm or working parameters near the operating location are automatically performed according to properties of the storm such as storm intensity, height, location and GIS information. From the analysis of 182 hail cases during the 31 hail weather processes over Chongqing, Dalian of Liaoning and Sanmenxia of Henan from 2006 to 2014, the storm auto classification technology developed in current study can significantly increase (decrease) the hit (false alarm) rate of storm tracking and recognition. The hit rate can reach 100%, and the false alarm rate is only 11.4%. The result suggests that this storm auto classification technology can enhance the automation of artificial hail prevention and contribute to the decision making for the operation of artificial hail suppression.